The long range goal of this research is to understand the roles(s) played by the superficial layers of the superior colliculus in visual function. An important aspect of this inquiry is to understand the spatial transformations in the retina's projection to the colliculus.l The retinotectal projection consists of multiple components, differing in ganglion-cell class and eye of origin. Although these components differ in spatiotopic organization, they appear to share a common retinotopic map at the colliculus. A central goal, then, is to learn what factors determine the metrics of the collicular map and what this implies for the representation of spatial information by the colliculus. To initiate this analysis, the proposed studies will focus on the crossed W-cell projection, which dominates the retinotectal pathway and offers unique technical advantages for such an inquiry. It is often assumed that ganglion cell density specifies central magnification (principle of peripheral scaling). But each of the ganglion cell classes projecting to the colliculus exhibits a unique retinal distribution so if retinotopic relationships are to be preserved, no more than a single ganglion cell class may strictly satisfy the peripheral scaling principle. Pilot data suggest that no class does. Another common assumption is that ganglion-cell receptive-field diameter is inversely proportional to ganglion-cell density so that a fixed number of ganglion cells "sees" any visual point (principle of uniform coverage factor). But a close examination of available evidence suggest that the major retinotectal ganglion cell type - the W-cell - may violate this relationship. The proposed studies will test directly whether tectally projecting W-cells do in fact violate these two mapping rules. A subsequent experiment will test an important implication of these apparent violations, namely, that the patch of W-cell afferents potentially activated by a visual point varies dramatically in size and shape over the collicular map. The findings will provide a basis for future work on collicular spatial transformations at the postsynaptic level. The issues explored here are of broad relevance because they arise also for the retinogeniculate projection and any other pathway in which afferent systems with disparate spatial organizations converge on a common central map. The principles involved also have important implications for the developmental mechanism generating central topographies.